AVIF vs. JPEG XL in 2026: A Practitioner's Honest Guide
AVIF and JPEG XL both claim to dethrone JPEG. Here's an expert breakdown of when to use each—and why the answer isn't as obvious as the benchmarks suggest.
TinyImage Team
Lead Architect
December 17, 2025
Published
6 min
Read time
Topics
Table of Contents
AVIF vs. JPEG XL in 2026: A Practitioner's Honest Guide
There's a format war being fought in your CDN logs, and most developers are picking sides based on benchmark charts that were out of date the moment they were published.
Let me save you the hours of reading: both formats are excellent, neither is universally superior, and you're probably asking the wrong question. Here's what actually matters.
Why This Choice Matters More Than You Think
A 150KB hero image on your homepage delivered over a 4G connection doesn't just slow down the page—it costs you Google ranking points. LCP (Largest Contentful Paint), the Core Web Vital that matters most for SEO, is exquisitely sensitive to image weight.
Choosing the right modern codec for the right job is the fastest, zero-infrastructure performance win available to any developer today.
AVIF: The Contender with a Video Pedigree
AVIF (AV1 Image File Format) was never designed to be an image format. It was engineered as a frame extractor from the AV1 video codec by the Alliance for Open Media—the same consortium that brought you Netflix-quality streaming at lower bitrates. That origin story explains both its greatest strength and its most persistent weakness.
What AVIF excels at
Compression efficiency at low bitrates is breathtaking. At quality settings below where WebP starts to pixelate, AVIF consistently delivers images that look 15–25% better to the human eye, measured by both SSIM and DMOS scores. For thumbnail grids where you're serving 80 to 200 images on a single e-commerce page, this is enormous.
Browser support is now a non-issue. As of early 2026, AVIF is supported natively in Chrome (since v85), Firefox (v93), and—crucially—Safari (v16). Combined, this covers over 95% of the global web-browsing market. The <picture> fallback to WebP or JPEG for the remaining fraction is trivially implemented.
Wide color gamut and HDR. AVIF supports up to 12-bit depth with BT.2020 and P3 color spaces. If you're serving lifestyle photography or product imagery where color accuracy is a brand concern, this matters.
Where AVIF falls short
Encoding speed. At maximum compression (useful for static build pipelines), AVIF encoding is 3–10x slower than WebP per image. In a live, user-facing compression tool running in the browser—like the engine TinyImage uses—WASM bridges this gap significantly, but it's worth knowing.
Tiling artifacts at extreme compression. AVIF divides images into tiles for parallel processing. At very aggressive settings (quality below 30), transitions between tiles can become visible in smooth gradients—like a sky or a soft-lit studio background. For product photography shot on white, this is often invisible. For lifestyle images with complex tonal transitions, it can be a problem.
JPEG XL: The Format Built for Photographers
JPEG XL (often written as JXL) emerged from a fundamentally different design philosophy. Rather than borrowing from video compression, its creators—engineers from Google and Cloudinary—asked a pointed question: what would JPEG look like if it were designed today?
The answer is a format with a few tricks that AVIF simply cannot match.
What JXL excels at
Lossless JPEG recompression. This is JXL's most unique capability, and it's genuinely mind-bending. JXL can take an existing JPEG file and transcode it into JXL format—losslessly—with a 20–22% file size reduction. You don't regenerate from the original; you simply re-encode the existing compressed data more efficiently. This makes JXL the only path to shrinking legacy JPEG archives without any generational quality loss.
Visual character at low bitrates. When AVIF's tile-based codec creates blocking artifacts, JXL's transform-based approach creates a different, far more forgiving failure mode: gentle blurring. Natural-looking degradation is a major UX upgrade in applications where bandwidth must be conserved—think news sites, mobile readers, and bandwidth-poor regions.
Progressive decoding. JXL was designed from the ground up to support progressive rendering—the JPEG trick of loading a blurry-then-sharp version of an image as data arrives. AVIF's architecture makes this difficult to implement. For above-the-fold hero images, JXL's progressive loading can make a page feel faster even if the bytes-on-the-wire are comparable.
Where JXL falls short
Browser adoption still lags AVIF. Though Microsoft Edge and Safari both formally support JXL, Chrome's support landed only behind a flag until 2025—a year of lost momentum that created legitimate gaps in older mobile browser versions. In 2026, you should always serve JXL with a full fallback chain.
The Practical Framework: When to Use Which
Stop thinking about this as AVIF versus JXL. Think of it as a context-aware decision:
| Use Case | Recommended Format | Why |
|---|---|---|
| E-commerce thumbnails (high volume) | AVIF | Smaller files at acceptable quality |
| Photography portfolio / e-commerce hero | JXL | Better fine detail, graceful blurring |
| Migrating legacy JPEG CDN | JXL | Lossless recompression, no sources needed |
| User-uploaded content (unknown type) | AVIF | Universal support is more reliable |
| Above-the-fold branding images | JXL | Progressive decode improves perceived speed |
| Batch-compress thousands of assets | AVIF | Faster encode pipelines |
The Privacy Angle No One Talks About
Most codec comparison articles assume you're uploading your images to a server-side conversion tool. That means sending potentially sensitive product shots, unreleased campaign imagery, or client work to someone else's infrastructure.
TinyImage's WASM engine runs the AVIF and JXL encoders directly in your browser. Judge the results for yourself in real-time without your source files ever leaving your machine. It's not just a developer convenience—it's a legitimate data security posture.
The Bottom Line
In 2026, the correct format strategy for a performance-conscious development team is:
- AVIF as your primary modern format for most web imagery.
- JXL as an intentional choice for photography-heavy use cases and any legacy JPEG migration.
- WebP as your baseline fallback where AVIF/JXL aren't supported.
- JPEG only as the final
<img src>fallback for truly ancient browsers.
The benchmarks will tell you one format wins. The real world will tell you the better question is: wins at what?
Deploy Visual Excellence
Put what you've learned into practice with TinyImage.Online - the free, privacy-focused image compression tool that works entirely in your browser.
